qpid-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Alan Conway <acon...@redhat.com>
Subject Re: Proton's road ahead
Date Mon, 31 Oct 2016 18:51:12 GMT
On Mon, 2016-10-31 at 18:26 +0000, Rob Godfrey wrote:
> On 31 October 2016 at 17:28, Robbie Gemmell <robbie.gemmell@gmail.com
> >
> wrote:
> 
> > > I was going to bring up a similar question - do we believe we are
> > actually
> > > 
> > > getting benefit from trying to keep the API of Proton-J and
> > > Proton-C
> > > "identical"?  It's been a while since I worked on it, but my
> > > feelings at
> > > the time were that the enforcing of API identity (rather than
> > equivalence)
> > > 
> > > actually made proton-j much less natural to use and not
> > > insignificantly
> > > more difficult to maintain.  I'd certainly like to see something
> > > with
> > > functional equivalence between C and Java, but it seems like
> > > we've
> > actually
> > > 
> > > fallen far behind on that with the lack of a reactive api.
> > > 
> > 
> > No, I don't think keeping API 'identical' has given benefit to
> > warrant
> > some of the less natural cases. That said, I'm not sure its
> > actually
> > tried as much as it was in the past either, since they have
> > actually
> > diverged in API in cases already, e.g. SASL handling is entirely
> > different between them these days (though partly because proton-c
> > changed and proton-j did not).
> > 
> 
> :-) Yeah - that was my impression too... My intent here was really to
> see
> if we can better restate the goals of Proton-J in particular so that
> we
> focus on an equivalence and maybe at the same time then look to keep
> the
> two libraries closer in functionality (i.e. not leaving the Java
> libraries
> behind because it is a pain in the ass to try to write C code in
> Java)

+1. Interop is the most important issue. Keeping things similar to
lower learning curves is a good thing, but forcing one language into
the mold of another raises the learning curve, so there's a balance to
be struck. I think "native X programmer can walk up and use the X
binding" should weigh a little more than "programmer who used Y binding
can immediately use the X binding". Both of those are far more
important than "it lets us re-use automated tests in Python and gloss
over writing tests in the binding languages."



---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@qpid.apache.org
For additional commands, e-mail: users-help@qpid.apache.org


Mime
View raw message